Taipei, Friday, Apr 26, 2024, 11:14

News

Taiwan’s AI Supercomputer “Taiwania 2” – Refreshes Performance Ranking Record

By Korbin Lan
Published: Nov 14,2018

TAIPEI, Taiwan - The Ministry of Science and Technology (MOST) announced that the National Applied Research Laboratories (RARLabs) National Center for High-Performance Computing (NCHC) in cooperation with the following three domestic companies: Quanta, Asus, and TaiwanMobile has built the Taiwania 2 artificial intelligence supercomputer host.

More on This

Taiwan’s ARRC Accomplishes Its First Rocket Launch in Pingtung County

PINGTUNG, Taiwan - The Pingtung County’s scientific research rocket launch site in Peony Township hold its first launc...

MOST: New Record on Ultrafast Green Micro-LED Array

“Ultrafast optical wireless data streaming for transmitting a high-definition video signal within one second with dire...

With an excellent performance of 9 PFLOPS, Taiwania 2 is ranked number 20 in the latest release of the world’s Top 500 major high-speed computer hosts, and in terms of energy efficiency it is ranked number 10 in the Green 500 making it the top scorer in all of Taiwan’s supercomputer history.

The Taiwania 2 supercomputer is comprised of 252 nodes, each of which is equipped with two CPUs and eight of the most advanced GPUs, and the host architecture design is synchronized with international trends. Therefore, when scientists make use of big data to carry out deep learning, the Taiwania 2 has improved performance.

In addition, in terms of energy conservation, Taiwania 2 features energy efficiency of up to 11.285 GF/W and consumes 798 KW of electricity when performing 9 PFLOPS of calculations making it the most energy-efficient high-speed computing host in Taiwan’s history.

Taiwania 2 is scheduled to be formally started up during the second half of next year, and in addition to providing rapid computing power, large storage space, and secure networking, through the creation of a cloud AI computing platform, Taiwania will become Taiwan’s largest data market and modeling market and provide industry and academic communities with more real-time and convenient computing services.

(TR/ Phil Sweeney)

CTIMES loves to interact with the global technology related companies and individuals, you can deliver your products information or share industrial intelligence. Please email us to en@ctimes.com.tw

2593 viewed

Most Popular

comments powered by Disqus